Junior Machine Learning Scientist

QARL AI; Montreal (administrative region); On-site; CAD 80,000 - 100,000; Part time

Job summary

A pioneering AI startup in Montreal is looking for a contractor to evaluate state-of-the-art Speech-to-Text solutions for their AI voice agent project. This role involves researching algorithms and creating testing frameworks, with a strong emphasis on transcription accuracy and integration support. The ideal candidate will have a Ph.D. in Computer Science and hands-on experience with relevant technologies, contributing to a dynamic team focused on cutting-edge AI solutions. Competitive compensation with potential full-time options and benefits is offered.

Benefits

Competitive compensation package

Lunch provided on Tuesdays

Learning stipends

Qualifications

• Track record of research in speech processing and generative AI.
• Hands-on experience with evaluation methodologies for STT systems.

Responsibilities

• Research and evaluate STT algorithms.
• Design testing pipelines for STT systems.
• Document findings and deliver recommendations.
• Provide technical integration support.

Skills

Speech-to-Text systems

Evaluation methodologies

Python

ML frameworks

Virtualization

Containerization

Cloud computing

Education

Doctorate (Ph.D.) in Computer Science or related fields

Tools

PyTorch

TensorFlow

Docker

Kubernetes

AWS

GCP

Azure

Job description

ABOUT THE COMPANY

QARL is an early‑stage startup building next‑generation interactive AI‑powered digital humans. Our mission is to reinvent how people engage with AI through fun, interactive video agents. Founded in 2024, we’re a small but mighty team led by serial entrepreneur and 3X exited founder Paul Kruszewski. Supported by leading partners including Google and AWS, we’re moving fast and pushing boundaries to define the future of human‑machine interaction.

OPPORTUNITY

In this 1‑month contract role, you’ll dive deep into researching, evaluating, and benchmarking state‑of‑the‑art Speech‑to‑Text (STT) solutions for our AI‑based voice agent pipeline. This is a rare opportunity to work on bleeding‑edge technology alongside experienced leadership and a top‑notch technical team. This position has the potential to extend to a full‑time position for the right candidate.

JOB DESCRIPTION

Primary responsibilities

• Lead a focused 1‑month project to research and evaluate open‑source and commercial‑friendly STT algorithms as alternatives to our current implementation.
• Design and implement standalone test harnesses and benchmarking pipelines for candidate STT systems.
• Evaluate candidates across key criteria: transcription accuracy (word error rate), latency, streaming support, hallucination robustness, noise robustness, and multi‑language support.
• Document findings and deliver a comprehensive recommendation report to facilitate decision‑making.
• Provide technical support to development team members in integrating the selected solution.

Education level

• Doctorate (Ph.D.) in Computer Science or related fields. Preferably with a track record of research in speech processing and generative AI.

The ideal candidate will have hands‑on experience with:

• Speech‑to‑Text systems and evaluation methodologies (WER, latency benchmarking).
• Python and ML frameworks (PyTorch, TensorFlow).
• Virtualization and containerization (Docker, Kubernetes, etc.).
• Cloud computing (AWS, GCP, Azure).

BENEFITS

• Competitive compensation package, with the possibility of converting to full‑time employment with stock options, health benefits, and more.
• Fully in‑person work culture (no remote work) with office downtown Montreal with lunch provided on Tuesdays.
• Stipends applicable toward continuous learning and subscriptions to Anthropic Claude AI.
• Work with one of the world’s best teams in artificial emotional intelligent and digital humans.